The deposition of insoluble protein aggregates, known as amyloid fibrils, in various tissues and organs is associated with a number of neurodegenerative diseases, including Alzheimer's, Huntington's and Parkinson's diseases, senile systemic amyloidosis and spongiform encephalopathies [(Volkova K D, Kovalska V B, Balanda A O, Vermeij R J, Subramaniam V, Slominskii Y L, Yarmoluk S M (2007), “Cyanine dye-protein interactions: looking for fluorescent probes for amyloid structures”. J. Biochem. Biophys. Methods 70: 727-733); and (Stefani M, Dobson C M (2003), “Protein aggregation and aggregate toxicity: new insights into protein folding, misfolding diseases and biological evolution”. J. Mol. Med. 81: 678-699)]. Fibrillar deposits with characteristics of amyloid are also formed by several other proteins unrelated to disease, including the whey protein beta-lactoglobulin (BLG). All amyloid fibers, independent of the protein from which they were formed, have very similar morphology: long and unbranched, a few nanometers in diameter, and they all exhibit a cross-beta X-ray diffraction pattern. The ability to form amyloid fibrils of structurally and functionally diverse proteins, some of which are not associated with amyloid-deposition diseases, suggests that this property is common to all polypeptides Such amyloid structures are also known to possess a binding affinity for certain dyes, notably, Thioflavin T and Congo Red dyes.
Many proteins are known to be only marginally stable in solution, undergoing conformational changes due to various stresses during purification, processing and storage [(Arakawa T, Philo J S, Ejima D, Sato H, Tsumoto K. (2007), “Aggregation analysis of therapeutic proteins, part 3”. Bioprocess International November (52-70). Such stresses may include elevated temperature, agitation and exposure to extremes of pH, ionic strength, or various interfaces (e.g., an air-liquid interface) and high protein concentration (as observed for some monoclonal antibody formulations). A wide variety of aggregates are encountered in biopharmaceutical samples, which range in size and physiochemical characteristics (e.g., solubility, reversibility). Protein aggregates span a broad size range, from small oligomers that are only a couple nanometers in length to insoluble micron-sized aggregates that extend to millions of monomeric units. Structurally altered proteins have an especially strong tendency to aggregate, often leading to their eventual precipitation. Irreversible aggregation is a major problem for the long-term storage and stability of therapeutic proteins and for their shipment and handling.
Mechanisms of Protein Aggregation
Aggregation is a major degradation pathway that needs to be characterized and controlled during the development of protein pharmaceuticals. In the bioprocessing arena, the mechanisms of protein aggregation are still not fully understood, despite the fact that aggregation is a major problem in therapeutic protein development (Arakawa T, Philo J S, Ejima D, Tsumoto K, Arisaka F (2006), “Aggregation analysis of therapeutic proteins, part 1”. Bioprocess International 4 (10): 32-42). One plausible mechanism is that aggregation is driven or catalyzed by the presence of a small amount of a contaminant which serves as a nucleation site. That contaminant could be a damaged form of the protein product itself, host cell proteins, or even nonprotein materials, such as leachates from the container or resin particles associated with purification of the protein.
If the contaminant is the damaged protein itself, then its aggregation may lead to soluble oligomers, which become larger aggregates, visible particulates, or insoluble precipitates. Such soluble oligomers, host-cell contaminants, or nonprotein materials may serve as a nucleus onto which native proteins assemble and are incorporated into larger aggregates. Damaged forms of a protein product can also arise from chemical modification (such as oxidation or deamidation) and from conformationally damaged forms arising from thermal stress, shear, or surface-induced denaturation. Minimizing protein aggregation thus requires ensuring both chemical and physical homogeneity; that is, chemically modified or conformationally altered proteins must be removed from the final product.
A second mechanism that often leads to protein aggregation is initiated by the partial unfolding of the native protein during its storage. Protein conformation is not rigid—the structure fluctuates around the time-averaged native structure to different extents depending upon environmental conditions. Some partially or fully unfolded protein molecules are always present at equilibrium in all protein solutions, but most such molecules simply refold to their native structure. These unfolded proteins may in some instances, however, aggregate with other such molecules or may be incorporated into an existing aggregate nucleus, eventually forming larger aggregates, as described above. Factors such as elevated temperature, shaking (shear and air-liquid interface stress), surface adsorption, and other physical or chemical stresses may facilitate partial unfolding of proteins, leading to the cascade of events that cause aggregation.
A third aggregation mechanism is reversible self-association of the native protein to form oligomers. According to the law of mass action, the content of such reversible aggregates will change with total protein concentration. The tendency of different proteins to associate reversibly with one another is highly variable, and the strength of that association typically varies significantly with solvent conditions, such as pH and ionic strength. In principle, these reversible oligomers will dissociate completely as the protein becomes highly diluted, for example, after delivery of a therapeutic protein in vivo. Consequently, this class of aggregates is generally less of a concern than irreversible aggregates. Such reversible oligomers can eventually become irreversible aggregates, however. Preventing accumulation of irreversible aggregates may thus require minimizing the reversible association as well. Further, reversible self-association of proteins can significantly alter overall pharmaceutical properties of product solutions, such as solution viscosity.
Detection of reversible aggregates can be an especially challenging task. As such, aggregates can dissociate after their dilution during attempts to measure them. Additionally, the results of any analysis method incorporating a separation process in the workflow may depend very much upon the kinetic rates of the reversible association-dissociation reactions as well as the equilibrium constants.
One consequence of the complexities of monitoring aggregate formation processes is the difficulty of linking the effect (presence of aggregates) to its underlying cause, particularly because the key damage may occur at a time or place quite separated from the observed consequence. One example arises during the large-scale production of therapeutic monoclonal antibodies (MAbs). Acid stability plays a major role in the aggregation of MAbs because the process for their purification usually involves both low-pH elution from protein-A affinity columns and acid-treatment for viral inactivation.
The exposure of MAbs to a low-pH environment can result in small but significant conformational changes that can additionally depend upon factors such as temperature, and solvent composition. While such partially unfolded MAbs may not aggregate at low pH, they may aggregate during subsequent manufacturing steps involving changes in pH or ionic strength. A larger conformational change at low pH generally leads to more aggregates upon increasing the pH. Typically, protein aggregate formation from the low-pH structure is not a fast process, but it does occur slowly from the association of damaged monomers that have not returned to their fully native structure. This and other types of protein aggregation phenomena may not manifest themselves until months after manufacturing a particular lot of protein or until later stages of the product development process. Regardless of the mechanism of aggregation, preventing aggregation problems requires sensitive and reliable technologies for quantitative determination of aggregate content and aggregate characteristics.
Since the earliest clinical applications of protein pharmaceuticals in medicine, aggregation problems have been implicated in adverse reactions in humans and other safety issues. In order to minimize such risks from therapeutic proteins in the clinic, formulations must be optimized to minimize aggregation during storage, handling, and shipping.
Analysis of Protein Aggregation
The analysis of protein aggregation can be formally classified into four experimental types [(Arakawa T, Philo J S, Ejima D, Tsumoto K, Arisaka F (2006), “Aggregation analysis of therapeutic proteins, part 1”. Bioprocess International 4(10): 32-42); (Arakawa T, Philo J S, Ejima D, Tsumoto K, Arisaka F (2007), “Aggregation analysis of therapeutic proteins, part 2”. Bioprocess International 5(4): 36-47); (Arakawa T, Philo J S, Ejima D, Sato H, Tsumoto K (2007), “Aggregation analysis of therapeutic proteins, part 3”. Bioprocess International 5(10): 52-70) (Krishnamurthy R, Sukumar M, Das T K, Lacher N A (2008), “Emerging analytical technologies for biothererapeutics development”. Bioprocess International 6(5): 32-42)].
The first type of protein aggregation analysis is the most conventional approach, wherein a small volume of sample is applied to a separation medium and forms a band or zone. As the band migrates through the medium, the proteins separate according to differences in size, electrophoretic charge, or mass. Gel electrophoresis, size exclusion chromatography (SEC), field flow fractionation (FFF), and the occasionally used band sedimentation technique belong to this class of methods. The movement of the band or zone in these methods is often monitored using absorbance or refractive index detection.
In the second type of analysis, the sample initially and uniformly fills a measurement cell. When an electrical or centrifugal driving force is then applied, the protein moves along the applied field, leaving a protein-depleted solvent, which creates a boundary between protein-free and protein-containing solution phases. The movement of this boundary over time is measured. This mode of separation is used in analytical ultracentrifugation-sedimentation velocity (AUC-SV) and moving-boundary electrophoresis.
The third type of analysis is a measurement of particle size with no physical separation. An example of this method is referred to as correlation spectroscopy and it measures the fluctuation of particles in solution due to Brownian motion (i.e., measures protein diffusion coefficients). Fluctuations of scattered light and of fluorescence intensity have been employed in this type of measurement. One of the most widely employed methods in this category is referred to as dynamic light scattering (DLS).
SEC is the most commonly implemented control method and has become an industry benchmark for quantification of protein aggregates. SEC is seen as a versatile technique for separation and quantification of protein aggregates because of its high precision, high throughput, ease of use, compatibility with a quality control (QC) environment, and in most cases ability to accurately quantify protein aggregates. In spite of these strengths, several concerns exist with the technique including: a potential loss of aggregates (especially multimers), interaction of samples with a column matrix, the required change of a sample buffer matrix to an SEC mobile phase, and the inherent requirement for dilution of samples. Additionally, perturbation of the distribution of protein aggregates under standard SEC methodological conditions is possible.
AUC-SV relies on hydrodynamic separation of various species in a heterogeneous protein mixture under strong centrifugal force. AUC-SV complements SEC in resolving and quantifying low levels of protein aggregates. The main advantages of AUC-SV are seen in its ability to detect and measure higher order aggregates (which may elute in the void volume of an SEC column) and to conduct these measurements without exposing samples to a column resin or SEC mobile phase. AUC-SV is considered an accurate method because it does not require standards or dissociate aggregates; thus it can be used as an orthogonal method to verify the accuracy of SEC results. AUC-SV suffers from lower precision than SEC, however. The practical aspects of AUC-SV that impact precision and accuracy are beginning to be understood better, and several recent studies have demonstrated the utility of AUC-SV to detect and quantify aggregates present at relatively low (˜1%) levels. Despite its advantages, AUC-SV is not yet readily amenable for use as a routine release test in the biotechnology industry because of issues related to low throughput, the need for specialized equipment, performance problems at high protein concentrations, the need for skilled practitioners of the method, and difficulty in validating data analysis software.
DLS uses the time-dependent fluctuations of a scattered-light signal to calculate the hydrodynamic diameter of protein aggregates and their relative proportions. This method is highly sensitive to large aggregates because the intensity of scattered light increases proportionally with molecular weight. As a result, very large aggregates (e.g., a 1,000-mer) present at trace levels (≤0.1%) can be detected with high sensitivity. If present, such aggregates would elute in the void volume of an SEC column or they may be filtered out. Although this method is ideal for detecting very low mass fractions of large aggregates, it cannot resolve species that are similar in size. At least a three- to five-fold difference in hydrodynamic diameter is required for resolving different species. DLS is also not amenable to use as a control method because it is semi-quantitative and very sensitive to dust or other extraneous particles. Results also depend on the algorithm used for data analysis, which is often proprietary to the manufacturer of a particular instrument.
As an orthogonal technique to SEC and AUC-SV, analytical field-flow fractionation (aFFF) has gained popularity in recent years for its ability to fractionate protein aggregates without a column. aFFF most commonly uses two fluid flows (“fields”) in a channel to achieve particle separation based upon molecular weight and hydrodynamic size (diffusion coefficient). Injected macromolecular species are held in place by a cross flow on a semi-permeable membrane while a perpendicular channel flow carries molecules forward based on their diffusion coefficient, thereby providing size-based fractionation. Because aFFF involves no column interactions, it is considered a gentler separation technique than SEC. Concerns regarding the interaction of aggregates with the membrane have yet to be completely addressed, however. aFFF can be coupled with different detectors including light scattering, refractive index, and ultraviolet (UV) detectors. When compared with SEC, the precision and limit of detection of aFFF is inferior in the high-molecular-weight range, because of increased baseline noise. Experimental conditions (e.g., cross-flow rate) for reasonable separations in one size range are also not generally applicable to other size ranges, making the technique cumbersome, especially when analyzing a broad range of masses. Along with other limitations, such as the need for specialized equipment and a skilled operator, and the difficulty in validating the method prevents the use of aFFF in applications for release and stability monitoring.
Resolution and the size range that can be evaluated in one particular analysis vary widely among the above mentioned techniques. SEC cannot handle a large range of sizes because the pore size or degree of polymerization of the resin must be adjusted to the size of the protein species. If a protein sample contains widely different sizes, many techniques are unsuitable for analyzing all sizes simultaneously. FFF and DLS can cover a very large range of sizes, but in the case of DLS, resolution is generally fairly poor, and FFF entails some trade-off between resolution and dynamic range. SV-AUC is intermediate in capability relative to FFF and DLS. The dynamic range of SV-AUC is fairly good, generally a factor of 100 or more in molecular weight at any particular rotor speed. The resolution of SV-AUC is generally not ideal for separating monomer from dimer, compared with the best SEC columns (especially for lower molecular weight proteins). SV-AUC is often much better, however, than SEC for resolving moderate size oligomers, (tetramers to decamers).
The cited analytical techniques also differ significantly with respect to their overall sensitivity, in other words, their ability to detect and quantify small percentages of irreversible aggregates. SEC, FFF, and SV-AUC are all capable of detecting aggregates at levels as low as ˜0.1% when they are well separated from other species. The quantification of species that elute from SEC or FFF is quite good, but aggregates can easily be lost during the separation process. Thus, SEC and FFF may provide good precision but poor accuracy. For SV-AUC, loss of protein aggregates to surfaces is usually not a problem, but accurate quantification of small oligomers (dimer-tetramer) at total levels of ˜2% or less is quite difficult.
The sensitivity of DLS increases linearly with the stoichiometry of the protein aggregate. DLS is for all practical purposes useless for detecting oligomers smaller than an octamer, because the technique cannot resolve such oligomers from monomeric species, and for those protein aggregate species that are resolved, the accuracy of the weight fractions is quite poor, typically plus or minus factors of two to ten. DLS exhibits excellent sensitivity, however, for very large aggregate species, which can often be detected at levels far below 0.01% by weight.
Overall, no single analytical technique is ideal for every protein or is optimal for analyzing the wide range of aggregation problems that can arise with protein pharmaceutical formulation. One important industry trend are recent requests from regulatory agencies that the protein aggregation analytical method used for lot release and/or formulation development. Typically, this means SEC which is cross-checked through one or more orthogonal approaches to ensure detection of all relevant protein aggregate species. Comparison of protein aggregate content using various technologies is thus an emerging topic of interest in biotechnology research.
Fluorescent Dyes and Protein Aggregation
In a fourth method of aggregate analysis, fluorescent dyes have been used to stain amyloidogenic material in histology, while insights into the prerequisites and kinetics of amyloid formation have been obtained by the in vitro analysis of this process using similar dyes [(Volkova K D, Kovalska V B, Balanda A O, Losytskyy My, Golub A G, Vermeij R J, Subramaniam V, Tolmachev O I, Yarmoluk S M (2008), “Specific fluorescent detection of fibrillar α-synuclein using mono- and trimethine cyanine dyes”. Bioorganic & Medicinal Chemistry 16:1452-1459); (Volkova K D, Kovalska V B, Balanda A O, Vermeij R J, Subramaniam V, Slominskii Y L, Yarmoluk S M (2007), “Cyanine dye-protein interactions: looking for fluorescent probes for amyloid structures”. J. Biochem. Biophys. Methods 70:727-733); (Volkova K D, Kovalska V B, Segers-Nolten G M, Veldhuis G, Subramaniam V, Slominskii Y L, Yarmoluk S M (2009), “Detection and characterization of protein aggregates by fluorescence microscopy”. Biotechnic & Histochemistry 84(2): 55-61); (Demeule B, Gurny R, Arvinte T (2007), “Explorations of the application of cyanine dyes for quantitative α-synuclein detection”. International Journal of Pharmaceutics 329: 37-45]. The fluorescent probes, Thioflavin T and Congo Red, have been the most frequently used dyes to detect the presence of amyloid deposits. Both the benzothiazole dye Thioflavin T and the symmetrical sulfonated azo dye Congo Red have been adapted to study the formation of amyloid fibrils in solution using the fluorescence properties of these molecules. The amyloid aggregates cause large enhancements in fluorescence of the dye thioflavin T, exhibit green-gold birefringence upon binding the dye Congo red, and cause a red-shift in the absorbance spectrum of Congo red. Amyloid fibril detection assays have suffered from several drawbacks, however, when using Thioflavin T, Congo Red and their derivatives. For instance, Congo Red can bind to native α-proteins such as citrate synthase and interleukin-2 [Khurana R, Uversky V N, Nielsen L, Fink A L (2001), “Is Congo Red an Amyloid-specific Dye”. J. Biol. Chem 276: 22715-22721]. As a consequence of its poor optical properties, the Congo Red derivative Chrysamine-G only weakly stains neuritic plaques and cerebrovascular amyloid in postmortem tissue [Klunk W E, Debnath M L, Koros A M, Pettegrew J W (1998) “Chrysamine-G, a lipophilic analogue of Congo Red, inhibits A beta-induced toxicity in PC12 cells.”. Life Sci. 63: 1807-1814]. Furthermore, the binding of dyes can influence the stability of amyloid aggregates, and the interplay with other components (for example, during testing of potential amyloid inhibitors) is unpredictable [Murakami K, Irie K, Morimoto A, Ohigashi H, Shindo M, Nagao M, Shimizu T, Shirasawa T (2003), “Neurotoxicity and Physicochemical Properties of Aβ Mutant Peptides from Cerebral Amyloid Angiopathy: IMPLICATION FOR THE PATHOGENESIS OF CEREBRAL AMYLOID ANGIOPATHY AND ALZHEIMER'S DISEASE”. J. Biol. Chem 278: 46179-46187]. Importantly, there exists a great variability among the different amyloid fibrils in their ability to bind Congo Red and Thioflavin T. Fluorescence intensity using Thioflavin T can vary depending upon the structure and morphology of the amyloid fibrils [Murakami K, Irie K, Morimoto A, Ohigashi H, Shindo M, Nagao M, Shimizu T, Shirasawa T (2003), “Neurotoxicity and Physicochemical Properties of Aβ Mutant Peptides from Cerebral Amyloid Angiopathy: IMPLICATION FOR THE PATHOGENESIS OF CEREBRAL AMYLOID ANGIOPATHY AND ALZHEIMER'S DISEASE”. J. Biol. Chem 278: 46179-46187]. Despite the widespread use of Thioflavin T, its application to amyloid quantification often generates inconsistent and inaccurate results. Variations in spectral properties caused by buffer conditions and protein-dye ratios result in poor reproducibility, complicating the use of Thioflavin T for quantitative assessment of fibril formation. In the absence of other more reliable assays, investigators have relied heavily upon Thioflavin T as a reporter probe for amyloid protein aggregation. A reliable method for amyloid quantification likely would be useful not only for detecting mature amyloid fibrils, but also for monitoring the kinetics of fibrillogenesis, which is essential for better understanding of the underlying biophysics and mechanism of the protein aggregation process. Furthermore, such an assay would be a tool for discovery and development of therapeutic compounds capable of blocking protein aggregation.
Thus the design of new dyes which can selectively interact with fibrillar amyloidogenic proteins is of substantial importance for basic research, and has a crucial practical significance for biotechnology and medicine. Dialkylamino-substituted monomethine cyanine T-284 and meso-ethyl-substituted trimethine cyanine SH-516 have demonstrated higher emission intensity and selectivity to aggregated α-synuclein (ASN) than the classic amyloid stain Thioflavin T; while the trimethinecyanines T-49 and SH-516 exhibit specifically increased fluorescence in the presence of fibrillar β-lactoglobulin (BLG) [Volkova K D, Kovalska V B, Balanda A O, Vermeij R J, Subramaniam V, Slominskii Y L, Yarmoluk S M (2007), “Cyanine dye-protein interactions: looking for fluorescent probes for amyloid structures”. J. Biochem. Biophys. Methods 70: 727-733]. These dyes demonstrated the same or higher emission intensity and selectivity to aggregated BLG as Thioflavin T. Recently, Nile Red dye has been used to detect antibody A aggregate, but it did not stain all types of protein aggregates, underscoring the need to several analytical methods in order to assess protein aggregation [Demeule B, Gurny R, Arvinte T (2007), “Detection and characterization of protein aggregates by fluorescence microscopy”. International Journal of Pharmaceutics 329: 37-45].
Optimization of Protein Formulations
Another potential application of a fluorescence based protein aggregate detection technique relates to pharmaceutical protein formulations [(Kim S, Antwerp W P V, Gross T M, Gulati P S (2004), “Methods of evaluating protein formulation stability and surfactant-stabilized insulin formulations derived there from”. U.S. Pat. No. 6,737,401 B2); [(Hsu C C, Nguyen H M, Wu S S (1993), “Reconstituteable lyophilized protein formulation”. U.S. Pat. No. 5,192,737); (Andya J, Cleland J L, Hsu C C, Lam X M, Overcashier D E, Shire S J, Yang J Y-F, Wu S S-Y (2004), “Protein formulation”. U.S. Pat. No. 6,685,940 B20); [(Ludvigsen S, Schein M, Boving T E G, Bonde C, Lilleore A, Engelund D K, Nielsen B R (2008), “Stable formulations of peptides”. US patent: application 2008/0125361 A1)]. The physical stability of pharmaceutical protein formulations is of great importance because there is always a time delay between production, protein formulation and its subsequent delivery to a patient. The physical stability of a protein formulation becomes even more critical when using drug delivery devices to dispense the protein formulation, such as infusion pumps and the like. When the delivery device is worn close to the body or implanted within the body, a patient's own body heat and body motion, plus turbulence generated in the delivery tubing and pump, impart a high level of thermo-mechanical stress to a protein formulation. In addition, infusion delivery devices expose the protein to hydrophobic interfaces in the delivery syringes and catheters. These interfacial interactions tend to destabilize the protein formulation by inducing denaturation of the native structure of the protein at these hydrophobic interfaces.
In an optimized protein formulation, the protein should remain stable for several years, maintaining the active conformation, even under unfavorable conditions that may occur during transport or storage. Protein formulation screening needs to be performed before the assessment of safety, toxicity, ADME (absorption distribution metabolism excretion), pharmacology and the testing of biological activity in animals. Currently, protein formulation in the pharmaceutical industry is generally a slow process and would benefit from fast formulation screening approaches that do not require overly complicated instrumentation techniques.
The formulation of protein drugs is a difficult and time-consuming process, mainly due to the structural complexity of proteins and the very specific physical and chemical properties they possess. Most protein formulations contain excipients which are added to stabilize protein structure, such as a particular buffer system, isotonic substances, metal ions, preservatives and one or more surfactants, with various concentration ranges to be tested. The conventional analytical methods usually require a long period of time to perform, typically twenty or more days, as well as manual intervention during this period. The development of new formulations is costly in terms of time and resources. Moreover, even for a known protein formulation, batch to batch quality control analysis is often less than optimal using the current state of the art methods. Therefore, a versatile, reliable, rapid and resource-efficient analytical method is desired for both developing novel protein formulations and identifying protein stability in quality control procedures. The ideal analytical method would be sensitive, accurate, and linear over a broad range, resistant to sample-matrix interference, capable of measuring all possible structural variants of a protein, and compatible with high throughput screening.
A high throughput screening (HTS) platform for optimization of protein formulation has been proposed based upon the use of multi-well microplates ([(Capelle Martinus A H, Gurny R, Arvinte T (2009), “A high throughput protein formulation platform: Case study of salmon calcitonin”. Pharmaceutical Research 26(1): 118-128). Basically, such an HTS platform was envisioned to consist of two components: (i) sample preparation and (ii) sample analysis. Sample preparation involves automated systems for dispensing the drug and the formulation ingredients in both liquid and powder form. The sample analysis involves specific methods developed for each protein to investigate physical and chemical properties of the formulations in the microplates.
The techniques that could be coupled with such an HTS platform include UV-Visible absorbance/turbidity, light scatter, fluorescence intensity, resonance energy transfer, fluorescence anisotropy, Raman spectroscopy, circular dichroism, Fourier transform infrared spectroscopy (FTIR), surface plasmon resonance and fluorescence lifetime. Ideally, however, the analysis technique should be specific, quantitative, robust, cost-effective, easily accessed, easy to use and informative (Avinte et al utilized several assays coupled with HTS to optimize a salmon calcitonin formulation: turbidity (absorbance at 350 nm), intrinsic tyrosine fluorescence, 1-anilino-naphthalene-8-sulfonate (ANS) fluorescence and Nile Red fluorescence. Addition of the dyes (Nile Red and ANS) were employed to examine protein conformational changes. Their findings were in accordance with the salmon calcitonin formulations that were patented and used commercially, lending credence to the concept that fluorescent probe-based approaches can be employed in protein formulation optimization activities. The use of several complementary analytical methods permits the selection of formulations using carefully designed assay criteria. The investigators found that in some cases, an increase in turbidity was observed without an increase in ANS or Nile Red fluorescence. In other formulations, an increase in fluorescence was detected without an increase in turbidity. This suggests that these dyes are not necessarily measuring the exact same biophysical phenomenon as the turbidity measurements. Measuring the fluorescence of at least two dyes in combination with turbidity and intrinsic fluorescence was, therefore, recommended.
Among these techniques, fluorescence detection from externally added dyes, which enhances fluorescence intensity upon interacting with misfolded or aggregated protein, is most attractive, because this technique requires minimum protein concentration due to its high sensitivity and simple implementation on a microplate reader.
Real time stability testing of a particular formulation may demonstrate no immediately apparent effect on physical or chemical stability. Accelerated stability testing can help, therefore, in facilitating the determination of the most suitable excipients and concentrations. Storage at different target temperatures (0-50° C.), illumination of samples, mechanical stress (i.e., agitation that simulates handling and transportation), multiple freeze-thaw cycles (mimicking frozen storage, freeze drying), oxygen purging, increased humidity and seeding are different ways to accelerate protein degradation.
High throughput spectroscopy is a fast and versatile method for initial screening of the physical stability of protein formulations. The microplate well-based platform could be enhanced with accelerated stress testing and methods to determine chemical stability, e.g., electrophoresis, HPLC, mass spectrometry. For instance, Thioflavin T has been used to select and optimize FDA-approved surfactant(s) in insulin formulations using magnetically stirring to accelerate insulin aggregation (U.S. Pat. No. 6,737,401 B2).
Thermal Shift Assay
Fluorescent dyes have been used to monitor protein stability by systematically varying the temperature of test samples, also known as the Thermofluor® technique [(Pantoliano M W, Rhind A W, Salemme F R (2000), “Microplate thermal shift assay for ligand development and multivariable protein chemistry optimization”. U.S. Pat. No. 6,020,141); (Matulis D, Kranz J K, Salemme F R, Todd M J (2005), “Thermodynamic stability of carbonic anhydrase: Measurements of binding affinity and stoichiometry using thermofluor”. Biochemistry 44: 5258-5266); (Mezzasalma T M, Kranz J K, Chan W, Struble G T, Schalk-Hihi C, Deckman I C, Springer B A, Todd M J (2007), “Enhancing recombinant protein quality and yield by protein stability profiling”. J. Biomolecular Screening 12(3): 418-428); [(Volkova K D, Kovalska V B, Balanda A O, Losytskyy My, Golub A G, Vermeij R J, Subramaniam V, Tolmachev O I, Yarmoluk S M (2008), “Specific fluorescent detection of fibrillar α-synuclein using mono- and trimethine cyanine dyes”. Bioorganic & Medicinal Chemistry 16: 1452-1459); (Ericsson U B, Hallberg B M, DeTitta G T, Dekker N, Nordlund P (2006), “Thermofluor-based high-throughput stability optimization of proteins for structural studies”. Analytical Chemistry 357: 289-298); (Todd M J, Cummings M D, Nelen M I (2005), “Affinity assays for decrypting protein targets of unknown function”. Drug Discovery Today 2 (3): 267-273)]. Protein stability can be altered by various additives including but not limited to excipients, salts, buffers, co-solvents, metal ions, preservatives, surfactants, and ligands. Protein stability can be shifted by various stresses, including elevated temperature, referred to as thermal shift, or chemical denaturants, such as urea, guanidine isocyanate or similar agents. A protein stability shift assay offers a wide spectrum of applications in the investigation of protein refolding conditions, optimization of recombinant protein expression/purification conditions, protein crystallization conditions, selection of ligand/drug/vaccine/diagnostic reagents and protein formulations.
The classic thermal shift technology utilizes the dye SYPRO® Orange and involves the use of a melting point device to raise the temperature stepwise [(Raibekas A A (2008), “Estimation of protein aggregation propensity with a melting point apparatus”. Anylytical Biochemistry, 380: 331-332). Thermal shift technology is coupled with aggregation detection technologies, such as light scattering technology or internal fluorescence from protein (such as tyrosine or tryptophan) to monitor protein aggregation and unfolding respectively. This type of technology usually requires a high protein concentration, therefore, it is not cost-effective. In addition, thermal shift technology cannot work effectively on formulations with low protein concentrations or finalize protein formulations which require a very low detection limit (typically ˜1-5% protein aggregates).
Fluorometric Screening Assay for Protein Disulfide Isomerase (PDI)
Protein disulfide isomerase (PDI, EC5.3.4.1) is a 57-kDa enzyme expressed at high levels in the endoplasmic reticulum (ER) of eukaryotic cells [(Ferrari D M, Soling H D (1999), “The protein disulfide-isomerase family: unraveling a string of folds”. Biochem. J. 339: 1-10)]. PDI was the first enzyme known to possess the disulfide isomerase activity and has been well characterized over the past three decades. In ER, PDI catalyzes both the oxidation and isomerization of disulfides of nascent polypeptides. Under the reducing condition of the cytoplasm, endosomes and cell surface, PDI catalyzes the reduction of protein disulfide bonds.
Folding catalysts such as PDI and peptidylprolyl isomerase accelerate slow chemical steps that accompany folding. Disulfide bond formation can occur quite rapidly, even before the completion of synthesis, but for some proteins disulfide bond formation is delayed and occurs post-translationally. PDI catalyzes disulfide formation and rearrangement by thiol/disulfide exchange during protein folding in the ER. As a member of the thioredoxin superfamily, which also includes homologs such as ERp57, PDIp, ERp72, PDIr and ERp5, PDI has two independent but non-equivalent active sites, with one positioned close to the C-terminus and another close to the N-terminus. Each site possesses two cysteine residues (CGHC) that cycle between the dithiol and disulfide oxidation states. The disulfide bond at the active site of PDI is a good oxidant that directly introduces a disulfide bond into protein substrates. The dithiol redox state is essential for catalyzing disulfide rearrangements. The necessity of having oxidized and reduced active sites for catalysis of different steps results in a redox optimum. Besides its major role in the processing and maturation of secretory proteins in ER, PDI and its homologs have been implicated in other important cellular processes. For example, cellular insulin degradation occurs in a sequential fashion with several identified steps. The initial degradative step occurs in endosomes with two or more cleavages in the B chain occurring. This is followed by reduction of disulfide bonds by PDI, or a related enzyme, generating an intact A chain and fragments of B chain. The insulin fragments are further cleaved by multiple proteolytic systems, such as the lysosomal degradation pathway.
PDI and its homologs also play roles in the processing and maturation of various secretory and cell surface proteins in the ER following their synthesis. Several in vitro studies have also suggested a chaperone function of PDI, that is to assist in protein folding or refolding. During ER stress, as for example during hypoxia in endothelial cells and astrocytes in the cerebral cortex, PDI is up-regulated. This indicates that PDI is involved in protecting cells under pathological or stressful conditions.
Besides ER, PDI also exists on many cell surfaces, such as endothelial cells, platelets, lymphocytes, hepatocytes, pancreatic cells and fibroblasts. For the reductive activity of plasma membrane, PDI is required for endocytosis of certain exogenous macromolecules. The cytotoxicity of diphtheria toxin is blocked by PDI inhibitors, which block the cleavage of the inter-chain disulfide bonds in the toxin. PDI also mediates reduction of disulfide bonds in human immunodeficiency virus envelope glycoprotein 120, which is essential for infectivity. PDI inhibitors can thus prevent virus entry into cells. Such functional activities make PDI and its homologs attractive drug targets.
Biochemical assays related to measuring PDI activity have been described. (1) ScRNase assay: PDI converts scrambled (inactive) RNase into native (active) RNase that further acts on its substrate. The reported sensitivity of the assay is in the micromolar range [Lyles M M, Gilbert H F (1991). “Catalysis of the oxidative folding of ribonuclease A by protein disulfide isomerase: dependence of the rate on the composition of the redox buffer”. Biochemistry 30(3): 613-619]. (2) The Insulin Turbidity Assay: PDI breaks the two disulfide bonds between the two insulin chains (A and B) that results in precipitation of the B chain. This precipitation can be monitored by measuring turbidity (absorbance at 620 nm), which in turn indicates PDI activity. Sensitivity of this assay is in the micromolar range [Lundstrom J, Holmgren A (1990), “Protein disulfide-isomerase is a substrate for thioredoxin reductase and has thioredoxin-like activity”. J. Biol. Chem. 265(16): 9114-9120]. Recently an end-point, high throughput screening assay of PDI isomerase activity based on enzyme-catalyzed reduction of insulin in the presence of dithiothreitol using hydrogen peroxide as a stop reagent has been developed [(Smith A M, Chan J, Oksenberg D, Urfer R, Wexler D S, O W A, Gao L, McAlorum A, Huang S (2004). “A high-throughput turbidometric assay for screening inhibitors of protein disulfide isomerase activity” (J. Biomolecular Screening 9 (7): 614-620); (Huang S, Oksenberg D, Urfer R (2005). “High-throughput turbidometric assay for screening inhibitors of protein disulfide isomerase activity” (U.S. Pat. No. 6,977,142 B2).
(3) The Di-E-GSSG assay: This is the fluorometric assay that can detect picomolar quantities of PDI and is, therefore, considered the most sensitive assay to date for detecting PDI activity. Di-E-GSSG has two eosin molecules attached to oxidized glutathione (GSSG). The proximity of eosin molecules leads to the quenching of its fluorescence. Upon breakage of the disulfide bond by PDI, however, fluorescence increases 70 fold [Raturi A, Mutus B (2007). “Characterization of redox state and reductase activity of protein disulfide isomerase under different redox environments using a sensitive fluorescent assay”. Free Radic. Biol. Med. 43(1): 62-70]. Certain common excipients can cause signal generation as well, such as 2-mercaptoethanol and dithiothreitol.
In view of the important functional activities of PDI and homologous enzymes, sensitive, real-time, high throughput methods that are time and cost-effective are highly desirable.
Chaperone/Anti-Chaperone Activity
A chaperone is a protein that can assist unfolded or incorrectly folded proteins to attain their native state by providing a microenvironment in which losses due to competing folding and aggregation reactions are reduced. ((Puig A, Gilbert H F (1994), “Protein disulfide isomerase exhibits chaperone and anti-chaperone activity in the oxidative refoding of lysozyme”. The Journal of Biological Chemistry 269(10): 7764-7771). Chaperones also mediate the reversibility of pathways leading to incorrectly folded structures. One of the major complications encountered in both in vitro and in vivo protein folding is aggregation resulting from the commonly encountered low solubility of the unfolded protein or different folding intermediates. The efficiency of folding depends upon how the unfolded protein partitions between pathways leading to aggregation and pathways leading to the native structure. In vivo, the partitioning between productive and non-productive folding pathways may be influenced by “foldases” and molecular chaperones. Foldases accelerate folding by catalyzing the slow chemical steps, such as disulfide bond formation and proline isomerization that may retard folding. Molecular chaperones do not appreciably accelerate folding but bind to nonnative proteins in a way that is thought to inhibit non-productive aggregation and misfolding. In order to prevent these improper interactions, chaperones must be present at concentrations that are stoichiometric with the newly synthesized proteins. Consequently, chaperones are often found at very high concentrations in the cell.
PDI is a very abundant protein within cells. Although primarily classified as a foldase, PDI has also been shown to possess chaperone or anti-chaperone activity (Puig and Gilbert, the Journal of Biological Chemistry (1994) 269: 7764-7771). PDI accelerates lysozyme folding, and at high concentration, it displays a chaperone-like activity that prevents lysozyme misfolding and aggregation. In addition, PDI also exhibits an unusual “anti-chaperone” activity. Under conditions that favor lysozyme aggregation, low concentrations of PDI greatly reduce the yield of native lysozyme and facilitate the formation of aggregates that are extensively cross-linked by intermolecular disulfide bonds. Similarly, PDI breaks the two disulfide bonds between two insulin chains (A and B) that results in precipitation of The B chain, thus serving as an “anti-chaperone in this case.” (Lundstrom J, Holmgren A (1990), “Protein disulfide-isomerase is a substrate for thioredoxin reductase and has thioredoxin-like activity”. J. Biol. Chem. 265 (16): 9114-20).
Alpha-crystallin, a major protein component of the mammalian lens of the eye, belongs to the heat shock protein (Hsp) family and acts as a molecular chaperone by preventing aggregation of target proteins (e.g. beta and gama-crystallins) under stress conditions through the formation of stable, soluble high-molecular mass complexes with them. Aggregation of BLG (beta-lactoglobulin) occurs mainly via intermolecular disulfide bond exchange. Upon heating, BLG aggregates, which can be accelerated by subjecting the protein to either an elevated pH or through the additional of DTT. α-crystallin prevents heat-induced BLG aggregation, acting as a chaperone in the absence of DTT; in the presence of DTT, however, this chaperone activity is less efficient due to faster aggregation of heated and reduced beta-lactoglobulin. Another Hsp protein, Hsp 27, protects myosin 51 from heat-induced aggregation, but not from thermal denaturation and ATPase inactivation.
Highly sensitive fluorescent probes useful to monitoring various protein functions relating to aggregation should assist in formulation optimization. Preferably, these probes should be applicable to a broad ranges of proteins and concentrations even in the presence of excipients, salts and buffers, providing sensitive limits of detection and excellent linear dynamic ranges.